Forecasting COVID-19 Vaccination Trends in Indonesia using Machine Learning

  • Ahmad Fauzan Aqil National Quemoy University, Taiwan
  • Hsi-Chieh Lee National Quemoy University, Taiwan
  • Sofi Ismarilla Wardani National Quemoy University, Taiwan


The ongoing COVID-19 pandemic requires much research to deal with this problem. Medical treatment has resulted in vaccine findings that work as an immune system to block the COVID-19 reaction process. However, many of these developments are still undergoing improvement and periodic testing to found better results for humans. Therefore, forecasting trends of the COVID-19 vaccine in Indonesia is carried out to regularly predict vaccines' effectiveness by adjusting conditions. This forecasting uses the time-series forecasting method by prioritizing a machine learning process in predicting probably future forecasts. Based on the highest vaccine used, we propose ARIMA and Facebook Prophet as machine learning models to predict vaccine trends in each country. The Prophet model results achieved an RMSE score of 0.176, which these results contained vaccines distributed in Indonesia. Besides that, the ARIMA model achieved an RMSE score of 0.453 using the same dataset. The results obtained from this method can be considered a policy for the government to deal with the effective use of vaccines according to future needs. As a further development, this research can be reviewed by paying attention to external aspects such as social and economic factors affecting the COVID-19 vaccination. The results obtained are more comprehensive and representative than this research based on conditions that provide policies for handling COVID-19.